Application Of VDAR Model And VaR Method In Portfolio Selection

The financial market has always played an important role in the operation of the world economic system,and the Chinese stock market has always been in a crucial and important strategic position in the development of the entire modern socialist market economic system.Investors always expect higher returns with lower risks in the stock market.At this time,research on investment portfolio-related issues emerges as the times require.The profit opportunity is fleeting,so the related research based on stock returns and risk prediction is also deepening and developing,hoping to help investors to grasp the dynamics of the stock market first,explore the economic trend,and conduct high-yield and low-yield through suitable investment portfolio models.Risk-based strategies are formulated to make efficient investment choices.The academic community has carried out various researches on the stock market.The financial stock time series is one of the important carriers of stock market data.The analysis and research of the financial stock time series is helpful to grasp the stock market conditions.We know that the stock market presents the characteristics of complexity and nonlinearity,and the linear time series model cannot predict the stock data very well,and the investment portfolio model obtained by predicting the data has a low accuracy rate.Therefore,this paper selects the vector double autoregressive model(VDAR)to predict the stock market.Nonlinearity and correlation of changes in observed variables.At the same time,this paper establishes the VDARVa R model based on the VDAR model,which inherits the characteristics of the VDAR model,and the value at risk estimated by this model can still capture the nonlinear and correlation characteristics of the stock market.Therefore,the investment portfolio model constructed in this way can fully reflect the characteristics of the stock market and obtain a more appropriate investment strategy.In this paper,based on the return data of six stocks from January 2015 to January 2018,the theoretical analysis of the VDAR model such as parameter estimation and model prediction is carried out,and based on the VDAR model,the stock at risk Va R is estimated,and the Va R is used to estimate the value at risk.Portfolio selection models built by risk measurement tools.First,use factor analysis to reduce the dimensionality of the stock,and then establish a nonlinear vector double autoregression model(VDAR)according to different classifications,and use the maximum likelihood method to estimate its parameters to obtain a forecast model and obtain the expected return.Compared with the classical model VAR,it is found that its fitting effect and predictive ability are better.At the same time,the risk Va R value of the stock is calculated based on the VDAR model,and the Va R estimation is found to be valid through the ER ratio verification.Then,based on the rate of return and value at risk predicted by the model,a minimum variance portfolio model was established,and the corresponding portfolio strategy was obtained,which was applied to the real stock market and verified that the strategy model can maintain a good profit for a period of time.ability.Finally,this paper uses the ant colony algorithm to optimize the multi-objective investment portfolio strategy model,mainly based on the optimization method of transition probability and the optimization method based on the sparse distance of adjacent points to find the optimal investment portfolio,and it is obtained that it is composed of multiple solutions.An efficient frontier can satisfy investors with different risk preferences to choose appropriate portfolio strategies.The innovation of this paper is the application of nonlinear vector model to the stock market,which can not only reflect the nonlinear characteristics of the stock market but also show the mutual influence of financial market assets.Next,the VDAR-VAR model is established to estimate the Va R of stock,which can model the nonlinear of conditional expectation and conditional variance.Finally,in the formulation of portfolio strategy,not only the asset-normal method is used to calculate portfolio Va R value to estimate the risk of the model,but also the ant colony algorithm is used to optimize the portfolio model and find the optimal way suitable for the model.